Opinion
How maladaptation can structure biodiversity: eco-evolutionary island biogeography Timothy E. Farkas1*, Andrew P. Hendry2, Patrik Nosil1, and Andrew P. Beckerman1 1 2
Animal and Plant Sciences, University of Sheffield, Sheffield, S10 2TN, UK Redpath Museum, Biology, McGill University, Montreal, H3A 2K6, Canada
Current research on eco-evolutionary dynamics is mainly concerned with understanding the role of rapid (or ‘contemporary’) evolution in structuring ecological patterns. We argue that the current eco-evolutionary research program, which focuses largely on natural selection, should be expanded to more explicitly consider other evolutionary processes such as gene flow. Because multiple evolutionary processes interact to generate quantitative variation in the degree of local maladaptation, we focus on how studying the ecological effects of maladaptation will lead to a more comprehensive view of how evolution can influence ecology. We explore how maladaptation can influence ecology through the lens of island biogeography theory, which yields some novel predictions, such as patch isolation increasing species richness. Evolution and biodiversity: past research and a new direction The idea that evolution generates biodiversity through speciation [1,2] has been well recognized for over a century. More recently, ecologists and evolutionary biologists are learning how rapid (or ‘contemporary’) evolution can influence a variety of ecological processes over short timescales [3,4], and thus might be an important driver of biodiversity without causing speciation [5–10]. The past two decades have seen an accumulation of research supporting this idea, both in the field of community genetics (see Glossary) [11,12] and more recently in the field of eco-evolutionary dynamics, the latter focusing explicitly on how rapid evolution can influence populations, communities, and ecosystems on contemporary timescales [13–16]. Much current research in eco-evolutionary dynamics focuses on scenarios where natural selection causes local adaptation to divergent habitats (Figure 1A) and thereby generates divergent ecological dynamics. This approach generally invokes local adaptation as an almost inevitable consequence of divergent ecological pressures, and thereby Corresponding author: Farkas, T.E. ([email protected]). Keywords: maladaptation; biodiversity; eco-evolutionary dynamics; island biogeography; species interactions; gene flow. * Current address: Ecology and Evolutionary Biology, University of Connecticut, Storrs, CT 06269, USA. 0169-5347/ ß 2015 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tree.2015.01.002
154
Trends in Ecology & Evolution, March 2015, Vol. 30, No. 3
underemphasizes an important and fundamental aspect of evolutionary ecology. Namely, the degree of local adaptation varies on a continuum from highly adapted to poorly adapted, with some degree of maladaptation being common in nature [17]. On the one hand, simply recognizing the existence of maladaptation is trivial because local adaptation can rarely if ever be perfect [18]. On the other hand, quantifying the degree of maladaptation – that is, the location of a population along an adaptation continuum – is decidedly nontrivial because it can vary substantially through time or across space and is expected to have a diversity of consequences. Quantitatively, adaptation and maladaptation can be measured in percentages or proportions and are converse quantities (1 adaptation = maladaptation). Therefore, a population that is 100% locally adapted (achieving maximum possible fitness) is 0% maladapted, and vice versa. This formulation works very generally in theory, although in practice requires attention to nuances, such as the use of traits versus fitness to define maladaptation, and whether
Glossary Community genetics: a field of study investigating the influence of genetic variation within species on the ecology of communities and ecosystems. Connectivity: measure of predicted immigration to a habitat patch from surrounding patches, calculated using distances between patches, dispersal ability of the focal organism, and population sizes (see also Isolation). Eco-evolutionary dynamics: related to community genetics, a field of study investigating interactions between ecology and evolution on contemporary timescales. Founder effects: a stochastic process by which the genetic makeup of a population in a newly colonized habitat differs from that of the source population (see also Genetic drift). Gene flow: movement of alleles between populations via dispersal/migration, yielding modified allele frequency in recipient populations. Genetic drift: stochastic changes in allele frequency owing to finite population size. Inbreeding: increased population homozygosity due to mating among close relatives. Isolation: the inverse of connectivity, although traditionally with respect only to distance between island and mainland. Linkage disequilibrium: non-random association of alleles at multiple genomic loci (Mal)adaptation: a continuously variable spectrum of local adaptation, from very poorly adapted to very well adapted. Metacommunity: a group of local communities linked by dispersal. Metapopulation: a group of local populations linked by dispersal. Pleiotropy: when variants (alleles) of a single gene influence multiple phenotypic traits.
Opinion
Trends in Ecology & Evolution March 2015, Vol. 30, No. 3
(A)
Strong LA
Strong LA
Strong LA
Strong LA
Medium LA
Weak LA
(B)
TRENDS in Ecology & Evolution
Figure 1. Divergent local adaptation versus degrees of (mal)adaptation. Differentcolored trees represent divergent habitat types and therefore different selective environments. Moths represent an evolutionarily dynamic species, with colormatching to habitats representing local adaptation (LA). (A) Local adaptation to divergent habitats results in phenotypically divergent populations that each exhibit 100% local adaptation (0% maladaptation). (B) Evolutionary processes lead to variation in the degree of local adaptation, and hence to phenotypically divergent populations, but habitats all have the same type (i.e., there is no divergent selection).
maladaptation should be considered in an absolute or relative sense [18]. Our main thesis is that the degree of maladaptation in a given population can substantially influence aspects of local biodiversity (Figure 1B). To develop this argument, we outline three basic steps that link rapid evolution to biodiversity via maladaptation (Figure 2). In developing this argument we use the term ‘(mal)adaptation’ to emphasize the continuum [18], and the terms ‘maladaptation’ and ‘adaptation’ when we refer to relative locations on, or directional movement across, that continuum. In step 1, we highlight several mechanisms that can generate maladaptation, such as gene flow between populations adapted to divergent environments [17]. In step 2, we highlight how (mal)adaptation in one species can alter the abundances of other species in a community (‘species abundances’), according to theory on networks of ecological interactions, such as food webs [19,20]. In step 3, we emphasize how variation in species abundances arising from steps 1 and 2 can influence species richness through effects on extinction and colonization [21,22]. We then propose an eco-evolutionary theory of island biogeography [23,24] that incorporates (mal)adaptation, and suggest a research program for its study. We use island biogeography as a predictive framework, as opposed to parallel concepts such as the metacommunity [25,26], because island biogeography is a well-known and intuitive simplification of natural ecological complexity. Moreover, it focuses explicitly on the roles of island isolation and area as factors that influence rates of extinction and colonization, and are also factors that can influence (mal)adaptation. Island biogeography thus provides an excellent starting point for extending our understanding of evolutionary mechanisms that drive patterns of biodiversity, and we anticipate that future work could allow a similar extension to be made to metacommunity theory.
Step 1: evolution drives (mal)adaptation We provide a brief overview of how the four core evolutionary processes (genetic drift, gene flow, natural selection, and mutation) can generate quantitative variation in (mal)adaptation. We describe the basic ways in which (mal)adaptation is shaped by each mechanism, which often oppose one another, as well as relevant population demographic and geographic considerations, which become important for our arguments around island biogeography. Throughout this section we make the reasonable assumption that patch area and population size are positively related [27], and therefore imply effects of patch area when discussing effects of population size. Genetic drift Genetic drift is a stochastic process that can generate maladaptation by causing the spread and fixation of deleterious alleles. Although drift is expected to be strongest for neutral loci, it can also influence loci under selection, potentially compromising local adaptation. Similarly, founder effects can be expected to generate (mal)adaptation in newly colonized habitats. In addition, genetic drift and/or founder effects coupled with inbreeding can generate inbreeding depression, which can be manifested as generalized maladaptation unrelated to any particular ecological environment [28]. Genetic drift is most pronounced in small and isolated populations, and where gene flow from divergent populations is less likely (see also the next section). Thus, maladaptation arising from genetic drift will be greater, or more likely, in small and isolated populations than in large and well-connected populations. Gene flow Gene flow can both increase and decrease maladaptation, depending on whether migrants (dispersers) are locally adapted to different or similar environments. In the first case, gene flow is likely to be deleterious and cause maladaptation when metapopulations inhabit heterogeneous patch networks, wherein different alleles are favored in different environments [29–32]. In the second case, gene flow is likely to be beneficial in metapopulations inhabiting homogeneous patch networks, where it provides the raw material for adaptation [33,34] and can mitigate the depressive effects of inbreeding [28]. Importantly, these beneficial effects usually accrue at very low levels of gene flow, whereas deleterious effects are expected with higher levels of gene flow [35]. Given data on population sizes, patch spatial configuration, and patch-level phenotypes, it is possible to quantify the amount of gene flow from similar versus different environments, and thereby to predict expected levels of (mal)adaptation [31]. Natural selection and constraints on adaptation Although the primary outcome of natural selection is to increase adaptation, several types of constraints can interact with natural selection to generate maladaptation [36]. First, genetic correlations among traits caused by pleiotropy or linkage disequilibrium [37,38] can result in maladaptation of particular traits, even while other traits are becoming better adapted. Second, evolutionary conflict can cause maladaptation. For example, sexual conflict can 155
Opinion
Trends in Ecology & Evolution March 2015, Vol. 30, No. 3
Evoluonary processes Populaon level
1
(Mal)adaptaon
Populaon density
Populaon trait mean
2
Community level
Species abundances
Exncon/ colonizaon
3
Species richness TRENDS in Ecology & Evolution
Figure 2. Schematic of the processes by which (mal)adaptation influences species richness. Evolutionary processes (selection, gene flow, drift) lead to (mal)adaptation in a focal species. (Mal)adaptation can influence speciesinteraction traits and/or population density of that species, either of which can then influence species abundances in a local community. Changes in the abundances of species in the community can influence extinction/colonization dynamics throughout the community, resulting in modified local diversity. Note that the arrows are drawn unidirectionally to highlight the mechanism of focus, but are not intended to indicate that effects may not flow up this chain of events. Numbers refer to the three main steps of the process as outlined in the main text, and annotations indicate the level of biological organization at which each process/concept focuses.
cause maladaptation in a trait for one sex that is driven by selection on that trait in the other sex [39,40]. Third, environmental change can generate temporally variable selection that contributes to maladaptation [41,42]. Although patch connectivity and area have been investigated as important consequences for divergent selection in the context of adaptive radiation [43,44], future research will be necessary to determine whether connectivity and area can have direct consequences for selection generally. Mutation Mutation is the ultimate source of new genetic variation. Mutation can thus generate maladaptation directly through the production of deleterious alleles. These negative effects will be stronger in smaller populations, which are more prone to chance fixation of deleterious alleles [45,46]. Mutation is also a source of novel beneficial alleles, and the rate of their production can limit adaptation, particularly in small populations ([46], but see [47]). Step 2: (mal)adaptation can influence species abundances We outline here how (mal)adaptation within one species can influence the abundance of other species in a local 156
community (‘species abundances’) through species interactions. To begin, we emphasize that (mal)adaptation refers to variation in the fit of traits to the environment of organisms expressing them. However, for (mal)adaptation to influence species interactions, those traits must either be important to species interactions themselves (‘speciesinteraction traits’) or directly cause changes in population density of the focal species (Figure 3). Although changes in species-interaction traits can often be concurrent with changes in population density, this is not inevitably so. For instance, densities can change due to (mal)adaptation in traits not related to species interactions, such as temperature tolerance (Figure 3, ‘other traits’), and speciesinteraction traits can change without altering population density, as in the case of soft selection [48,49]. Following established concepts in food web ecology [19,20], the effects of (mal)adaptation on species-interaction traits and/or population density can propagate through entire species-interaction networks, influencing the population density and/or species-interaction traits of other species (Figure 3; ‘Species 2’). Expanding Figure 3 to a large network of interacting species (not pictured), (mal)adaptation can drive variation in species abundances for entire communities. To demonstrate the various ways in which (mal)adaptation can influence species abundances, we offer a three trophic level predator–herbivore–plant system as an example. In this example, inbreeding-based (mal)adaptation in a predator, causing reduced population density of the predator, might result in a higher abundance of the herbivore species on which it feeds. Alternatively (or additionally), gene flow between predator populations might lead to suboptimal predator foraging traits, which could similarly result in a higher abundance of the herbivore species. Furthermore, these direct effects of (mal)adaptation on adjacent species (predator–herbivore) might affect nonadjacent members through indirect interactions (predator–plant). Continuing the example above, maladaptation in a predator that increases herbivore abundance might cause a trophic cascade, thereby lowering the abundance of plant species on which they feed [50]. Step 3: (mal)adaptation can influence extinction and colonization We draw here on recent theoretical work which combines food-web theory with the classic equilibrium theory of island biogeography [23,24], to illustrate mechanisms through which (mal)adaptation that changes species abundances can influence species richness. In classical island biogeography, increasing island area and increasing connectivity to the mainland both increase colonization rates and decrease extinction rates, yielding higher equilibrium species richness (Figure 4A) [23]. In the ‘trophic theory of island biogeography’, extinction/colonization rates are additionally modified by food-web structure and species abundances, shifting predictions of equilibrium species richness [21,22]. For example, the likelihood of a predator successfully colonizing and persisting on an island should increase with the abundance its prey species, and should decrease with the abundance of its competitors (although the likelihood of
Opinion
Trends in Ecology & Evolution March 2015, Vol. 30, No. 3
Focal species
Populaon density Other traits Speciesinteracon traits
Species 2
Populaon density
Speciesinteracon traits
Other traits
(Mal)adaptaon TRENDS in Ecology & Evolution
Figure 3. Conceptual diagram of step 2: how (mal)adaptation can influence species abundances. (Mal)adaptation, here specifically the process by which evolution moves a population along the (mal)adaptation continuum, influences the traits of a focal species. Those traits might be related to species interactions (‘speciesinteraction traits’) or related to other aspects of the environment (‘other traits’). Both species-interaction traits and other traits might influence the population density of the focal species. Effects of (mal)adaptation on the focal species can propagate through a species interaction network because population density and/ or species-interaction traits in the focal species influence those of other species. Note: all arrows are unidirectional, but many species properties might well be drawn with bidirectional arrows.
an individual propagule arriving should be unaffected by resident community structure). Using knowledge of interaction strengths between the predator and its competitors, together with the abundances of each interacting species, it is possible to predict the colonization and extinction rates for the predator above and beyond those predicted using classical island biogeography. In step 2, we showed how density- or trait-based (mal)adaptation can influence species abundances through welldefined pathways in species interaction networks. Because the trophic theory of island biogeography demonstrates how variation in species abundances can modify extinction/ colonization dynamics, the trophic theory supports the notion that (mal)adaptation can influence colonization and extinction rates as well. Hence, (mal)adaptation can influence species richness. For example, a maladapted predator with inefficient foraging might increase the species richness of an assemblage on which it preys by increasing the likelihood of successful prey colonization, and by decreasing the likelihood of prey extinction. Eco-evolutionary island biogeography Integrating the above steps, we now outline a framework for eco-evolutionary island biogeography that incorporates (mal)adaptation. In classical island biogeography, connectivity and area can influence species richness through nicheneutral and stochastic effects on colonization/extinction dynamics. In addition, connectivity and area can influence (mal)adaptation through their effects on a variety of evolutionary processes (step 1). Thus, combining steps 1–3, connectivity and area can influence species richness through their effects on (mal)adaptation. In short, there are two independent pathways by which connectivity and area can influence species richness, both mediated by effects on colonization/extinction dynamics. We argue that predictions of equilibrium species richness will be most accurate
when considering an eco-evolutionary island biogeography that integrates the traditionally considered effects of patch connectivity and area (Figure 4A) with their effects via (mal)adaptation (Figure 4B,C). In any given metacommunity, the traditional and (mal)adaptation-mediated effects of connectivity and area on species richness might operate alone or in concert. If both pathways are operating simultaneously, they very well might have effects on species richness of contrasting magnitude, and might even have effects in contrasting directions. For example, in any scenario where connectivity increases maladaptive gene flow, and maladaptation leads to reduced species richness, traditional and (mal)adaptation-mediated effects of connectivity on species richness should oppose one another. We therefore argue that it is valuable (and possibly essential) to evaluate the effects of both mechanisms because eco-evolutionary island biogeography can yield predictions of species richness that equal, exacerbate (Figure 4B), nullify, or even invert (Figure 4C) traditionally recognized effects of patch connectivity and area. Two empirical examples from recent eco-evolutionary research highlight these effects and support their general application. Timema cristinae stick insects Gene flow between T. cristinae populations locally adapted to different host-plant species causes populations to become poorly camouflaged (maladapted) on their resident host plants [31]. Thus, connectivity among populations on different host-plant species strongly influences the degree of camouflage maladaptation [51]. Recent experiments show that camouflage maladaptation has the ecological consequence of decreasing the species richness of cohabitating arthropods, such as caterpillars and beetles, because avian predators that eat stick insects are attracted to host-plant patches harboring maladapted Timema [9]. In this scenario, the (mal)adaptation-mediated effect of connectivity is to decrease equilibrium species richness, opposing the traditionally predicted effect of connectivity (Figure 4A). If the (mal)adaptation-mediated effect outweighs the traditional effect, equilibrium arthropod species richness could actually be lower on well-connected patches in nature (Figure 4C). Further landscape-scale experimentation and observation will be necessary to determine whether natural patterns reflect the findings of the experiments, and to compare the magnitude and direction of the multiple effects of connectivity. Threespine stickleback Gene flow between lake and stream populations of threespine stickleback (Gasterosteus aculeatus) causes maladaptation in foraging morphology [52], and some evidence suggests it might reduce stickleback population density [53]. In a stream habitat, where well-adapted stickleback feed on benthic macroinvertebates, maladaptive gene flow from lake populations should relax predation pressure by decreasing foraging efficiency on the benthos. This effect could increase successful macroinvertebrate colonization from nearby streams and reduce the likelihood of extinction, yielding higher equilibrium species richness for this guild. 157
Opinion
Trends in Ecology & Evolution March 2015, Vol. 30, No. 3
(A)
Rate of immigraon or exncon
Key:
Ex
nc
Isolated (I) Connected (C)
Im
m
on
igr
a
on
Species richness I
(B)
Isolated, classical Connected, classical Isolated, eco-evoluonary Connected, eco-evoluonary
Ex
on
n
a
c
igr
on
m
Im
Rate of immigraon or exncon
Key:
< C
Species richness Classical: Eco-evoluonary:
(C)
I I
C C
Isolated, classical Connected, classical Isolated, eco-evoluonary Connected, eco-evoluonary
Ex
on
n
c
a igr
on
m
Im
Rate of immigraon or exncon
Key:
<
How maladaptation can structure biodiversity: eco-evolutionary island biogeography Timothy E. Farkas1*, Andrew P. Hendry2, Patrik Nosil1, and Andrew P. Beckerman1 1 2
Animal and Plant Sciences, University of Sheffield, Sheffield, S10 2TN, UK Redpath Museum, Biology, McGill University, Montreal, H3A 2K6, Canada
Current research on eco-evolutionary dynamics is mainly concerned with understanding the role of rapid (or ‘contemporary’) evolution in structuring ecological patterns. We argue that the current eco-evolutionary research program, which focuses largely on natural selection, should be expanded to more explicitly consider other evolutionary processes such as gene flow. Because multiple evolutionary processes interact to generate quantitative variation in the degree of local maladaptation, we focus on how studying the ecological effects of maladaptation will lead to a more comprehensive view of how evolution can influence ecology. We explore how maladaptation can influence ecology through the lens of island biogeography theory, which yields some novel predictions, such as patch isolation increasing species richness. Evolution and biodiversity: past research and a new direction The idea that evolution generates biodiversity through speciation [1,2] has been well recognized for over a century. More recently, ecologists and evolutionary biologists are learning how rapid (or ‘contemporary’) evolution can influence a variety of ecological processes over short timescales [3,4], and thus might be an important driver of biodiversity without causing speciation [5–10]. The past two decades have seen an accumulation of research supporting this idea, both in the field of community genetics (see Glossary) [11,12] and more recently in the field of eco-evolutionary dynamics, the latter focusing explicitly on how rapid evolution can influence populations, communities, and ecosystems on contemporary timescales [13–16]. Much current research in eco-evolutionary dynamics focuses on scenarios where natural selection causes local adaptation to divergent habitats (Figure 1A) and thereby generates divergent ecological dynamics. This approach generally invokes local adaptation as an almost inevitable consequence of divergent ecological pressures, and thereby Corresponding author: Farkas, T.E. ([email protected]). Keywords: maladaptation; biodiversity; eco-evolutionary dynamics; island biogeography; species interactions; gene flow. * Current address: Ecology and Evolutionary Biology, University of Connecticut, Storrs, CT 06269, USA. 0169-5347/ ß 2015 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.tree.2015.01.002
154
Trends in Ecology & Evolution, March 2015, Vol. 30, No. 3
underemphasizes an important and fundamental aspect of evolutionary ecology. Namely, the degree of local adaptation varies on a continuum from highly adapted to poorly adapted, with some degree of maladaptation being common in nature [17]. On the one hand, simply recognizing the existence of maladaptation is trivial because local adaptation can rarely if ever be perfect [18]. On the other hand, quantifying the degree of maladaptation – that is, the location of a population along an adaptation continuum – is decidedly nontrivial because it can vary substantially through time or across space and is expected to have a diversity of consequences. Quantitatively, adaptation and maladaptation can be measured in percentages or proportions and are converse quantities (1 adaptation = maladaptation). Therefore, a population that is 100% locally adapted (achieving maximum possible fitness) is 0% maladapted, and vice versa. This formulation works very generally in theory, although in practice requires attention to nuances, such as the use of traits versus fitness to define maladaptation, and whether
Glossary Community genetics: a field of study investigating the influence of genetic variation within species on the ecology of communities and ecosystems. Connectivity: measure of predicted immigration to a habitat patch from surrounding patches, calculated using distances between patches, dispersal ability of the focal organism, and population sizes (see also Isolation). Eco-evolutionary dynamics: related to community genetics, a field of study investigating interactions between ecology and evolution on contemporary timescales. Founder effects: a stochastic process by which the genetic makeup of a population in a newly colonized habitat differs from that of the source population (see also Genetic drift). Gene flow: movement of alleles between populations via dispersal/migration, yielding modified allele frequency in recipient populations. Genetic drift: stochastic changes in allele frequency owing to finite population size. Inbreeding: increased population homozygosity due to mating among close relatives. Isolation: the inverse of connectivity, although traditionally with respect only to distance between island and mainland. Linkage disequilibrium: non-random association of alleles at multiple genomic loci (Mal)adaptation: a continuously variable spectrum of local adaptation, from very poorly adapted to very well adapted. Metacommunity: a group of local communities linked by dispersal. Metapopulation: a group of local populations linked by dispersal. Pleiotropy: when variants (alleles) of a single gene influence multiple phenotypic traits.
Opinion
Trends in Ecology & Evolution March 2015, Vol. 30, No. 3
(A)
Strong LA
Strong LA
Strong LA
Strong LA
Medium LA
Weak LA
(B)
TRENDS in Ecology & Evolution
Figure 1. Divergent local adaptation versus degrees of (mal)adaptation. Differentcolored trees represent divergent habitat types and therefore different selective environments. Moths represent an evolutionarily dynamic species, with colormatching to habitats representing local adaptation (LA). (A) Local adaptation to divergent habitats results in phenotypically divergent populations that each exhibit 100% local adaptation (0% maladaptation). (B) Evolutionary processes lead to variation in the degree of local adaptation, and hence to phenotypically divergent populations, but habitats all have the same type (i.e., there is no divergent selection).
maladaptation should be considered in an absolute or relative sense [18]. Our main thesis is that the degree of maladaptation in a given population can substantially influence aspects of local biodiversity (Figure 1B). To develop this argument, we outline three basic steps that link rapid evolution to biodiversity via maladaptation (Figure 2). In developing this argument we use the term ‘(mal)adaptation’ to emphasize the continuum [18], and the terms ‘maladaptation’ and ‘adaptation’ when we refer to relative locations on, or directional movement across, that continuum. In step 1, we highlight several mechanisms that can generate maladaptation, such as gene flow between populations adapted to divergent environments [17]. In step 2, we highlight how (mal)adaptation in one species can alter the abundances of other species in a community (‘species abundances’), according to theory on networks of ecological interactions, such as food webs [19,20]. In step 3, we emphasize how variation in species abundances arising from steps 1 and 2 can influence species richness through effects on extinction and colonization [21,22]. We then propose an eco-evolutionary theory of island biogeography [23,24] that incorporates (mal)adaptation, and suggest a research program for its study. We use island biogeography as a predictive framework, as opposed to parallel concepts such as the metacommunity [25,26], because island biogeography is a well-known and intuitive simplification of natural ecological complexity. Moreover, it focuses explicitly on the roles of island isolation and area as factors that influence rates of extinction and colonization, and are also factors that can influence (mal)adaptation. Island biogeography thus provides an excellent starting point for extending our understanding of evolutionary mechanisms that drive patterns of biodiversity, and we anticipate that future work could allow a similar extension to be made to metacommunity theory.
Step 1: evolution drives (mal)adaptation We provide a brief overview of how the four core evolutionary processes (genetic drift, gene flow, natural selection, and mutation) can generate quantitative variation in (mal)adaptation. We describe the basic ways in which (mal)adaptation is shaped by each mechanism, which often oppose one another, as well as relevant population demographic and geographic considerations, which become important for our arguments around island biogeography. Throughout this section we make the reasonable assumption that patch area and population size are positively related [27], and therefore imply effects of patch area when discussing effects of population size. Genetic drift Genetic drift is a stochastic process that can generate maladaptation by causing the spread and fixation of deleterious alleles. Although drift is expected to be strongest for neutral loci, it can also influence loci under selection, potentially compromising local adaptation. Similarly, founder effects can be expected to generate (mal)adaptation in newly colonized habitats. In addition, genetic drift and/or founder effects coupled with inbreeding can generate inbreeding depression, which can be manifested as generalized maladaptation unrelated to any particular ecological environment [28]. Genetic drift is most pronounced in small and isolated populations, and where gene flow from divergent populations is less likely (see also the next section). Thus, maladaptation arising from genetic drift will be greater, or more likely, in small and isolated populations than in large and well-connected populations. Gene flow Gene flow can both increase and decrease maladaptation, depending on whether migrants (dispersers) are locally adapted to different or similar environments. In the first case, gene flow is likely to be deleterious and cause maladaptation when metapopulations inhabit heterogeneous patch networks, wherein different alleles are favored in different environments [29–32]. In the second case, gene flow is likely to be beneficial in metapopulations inhabiting homogeneous patch networks, where it provides the raw material for adaptation [33,34] and can mitigate the depressive effects of inbreeding [28]. Importantly, these beneficial effects usually accrue at very low levels of gene flow, whereas deleterious effects are expected with higher levels of gene flow [35]. Given data on population sizes, patch spatial configuration, and patch-level phenotypes, it is possible to quantify the amount of gene flow from similar versus different environments, and thereby to predict expected levels of (mal)adaptation [31]. Natural selection and constraints on adaptation Although the primary outcome of natural selection is to increase adaptation, several types of constraints can interact with natural selection to generate maladaptation [36]. First, genetic correlations among traits caused by pleiotropy or linkage disequilibrium [37,38] can result in maladaptation of particular traits, even while other traits are becoming better adapted. Second, evolutionary conflict can cause maladaptation. For example, sexual conflict can 155
Opinion
Trends in Ecology & Evolution March 2015, Vol. 30, No. 3
Evoluonary processes Populaon level
1
(Mal)adaptaon
Populaon density
Populaon trait mean
2
Community level
Species abundances
Exncon/ colonizaon
3
Species richness TRENDS in Ecology & Evolution
Figure 2. Schematic of the processes by which (mal)adaptation influences species richness. Evolutionary processes (selection, gene flow, drift) lead to (mal)adaptation in a focal species. (Mal)adaptation can influence speciesinteraction traits and/or population density of that species, either of which can then influence species abundances in a local community. Changes in the abundances of species in the community can influence extinction/colonization dynamics throughout the community, resulting in modified local diversity. Note that the arrows are drawn unidirectionally to highlight the mechanism of focus, but are not intended to indicate that effects may not flow up this chain of events. Numbers refer to the three main steps of the process as outlined in the main text, and annotations indicate the level of biological organization at which each process/concept focuses.
cause maladaptation in a trait for one sex that is driven by selection on that trait in the other sex [39,40]. Third, environmental change can generate temporally variable selection that contributes to maladaptation [41,42]. Although patch connectivity and area have been investigated as important consequences for divergent selection in the context of adaptive radiation [43,44], future research will be necessary to determine whether connectivity and area can have direct consequences for selection generally. Mutation Mutation is the ultimate source of new genetic variation. Mutation can thus generate maladaptation directly through the production of deleterious alleles. These negative effects will be stronger in smaller populations, which are more prone to chance fixation of deleterious alleles [45,46]. Mutation is also a source of novel beneficial alleles, and the rate of their production can limit adaptation, particularly in small populations ([46], but see [47]). Step 2: (mal)adaptation can influence species abundances We outline here how (mal)adaptation within one species can influence the abundance of other species in a local 156
community (‘species abundances’) through species interactions. To begin, we emphasize that (mal)adaptation refers to variation in the fit of traits to the environment of organisms expressing them. However, for (mal)adaptation to influence species interactions, those traits must either be important to species interactions themselves (‘speciesinteraction traits’) or directly cause changes in population density of the focal species (Figure 3). Although changes in species-interaction traits can often be concurrent with changes in population density, this is not inevitably so. For instance, densities can change due to (mal)adaptation in traits not related to species interactions, such as temperature tolerance (Figure 3, ‘other traits’), and speciesinteraction traits can change without altering population density, as in the case of soft selection [48,49]. Following established concepts in food web ecology [19,20], the effects of (mal)adaptation on species-interaction traits and/or population density can propagate through entire species-interaction networks, influencing the population density and/or species-interaction traits of other species (Figure 3; ‘Species 2’). Expanding Figure 3 to a large network of interacting species (not pictured), (mal)adaptation can drive variation in species abundances for entire communities. To demonstrate the various ways in which (mal)adaptation can influence species abundances, we offer a three trophic level predator–herbivore–plant system as an example. In this example, inbreeding-based (mal)adaptation in a predator, causing reduced population density of the predator, might result in a higher abundance of the herbivore species on which it feeds. Alternatively (or additionally), gene flow between predator populations might lead to suboptimal predator foraging traits, which could similarly result in a higher abundance of the herbivore species. Furthermore, these direct effects of (mal)adaptation on adjacent species (predator–herbivore) might affect nonadjacent members through indirect interactions (predator–plant). Continuing the example above, maladaptation in a predator that increases herbivore abundance might cause a trophic cascade, thereby lowering the abundance of plant species on which they feed [50]. Step 3: (mal)adaptation can influence extinction and colonization We draw here on recent theoretical work which combines food-web theory with the classic equilibrium theory of island biogeography [23,24], to illustrate mechanisms through which (mal)adaptation that changes species abundances can influence species richness. In classical island biogeography, increasing island area and increasing connectivity to the mainland both increase colonization rates and decrease extinction rates, yielding higher equilibrium species richness (Figure 4A) [23]. In the ‘trophic theory of island biogeography’, extinction/colonization rates are additionally modified by food-web structure and species abundances, shifting predictions of equilibrium species richness [21,22]. For example, the likelihood of a predator successfully colonizing and persisting on an island should increase with the abundance its prey species, and should decrease with the abundance of its competitors (although the likelihood of
Opinion
Trends in Ecology & Evolution March 2015, Vol. 30, No. 3
Focal species
Populaon density Other traits Speciesinteracon traits
Species 2
Populaon density
Speciesinteracon traits
Other traits
(Mal)adaptaon TRENDS in Ecology & Evolution
Figure 3. Conceptual diagram of step 2: how (mal)adaptation can influence species abundances. (Mal)adaptation, here specifically the process by which evolution moves a population along the (mal)adaptation continuum, influences the traits of a focal species. Those traits might be related to species interactions (‘speciesinteraction traits’) or related to other aspects of the environment (‘other traits’). Both species-interaction traits and other traits might influence the population density of the focal species. Effects of (mal)adaptation on the focal species can propagate through a species interaction network because population density and/ or species-interaction traits in the focal species influence those of other species. Note: all arrows are unidirectional, but many species properties might well be drawn with bidirectional arrows.
an individual propagule arriving should be unaffected by resident community structure). Using knowledge of interaction strengths between the predator and its competitors, together with the abundances of each interacting species, it is possible to predict the colonization and extinction rates for the predator above and beyond those predicted using classical island biogeography. In step 2, we showed how density- or trait-based (mal)adaptation can influence species abundances through welldefined pathways in species interaction networks. Because the trophic theory of island biogeography demonstrates how variation in species abundances can modify extinction/ colonization dynamics, the trophic theory supports the notion that (mal)adaptation can influence colonization and extinction rates as well. Hence, (mal)adaptation can influence species richness. For example, a maladapted predator with inefficient foraging might increase the species richness of an assemblage on which it preys by increasing the likelihood of successful prey colonization, and by decreasing the likelihood of prey extinction. Eco-evolutionary island biogeography Integrating the above steps, we now outline a framework for eco-evolutionary island biogeography that incorporates (mal)adaptation. In classical island biogeography, connectivity and area can influence species richness through nicheneutral and stochastic effects on colonization/extinction dynamics. In addition, connectivity and area can influence (mal)adaptation through their effects on a variety of evolutionary processes (step 1). Thus, combining steps 1–3, connectivity and area can influence species richness through their effects on (mal)adaptation. In short, there are two independent pathways by which connectivity and area can influence species richness, both mediated by effects on colonization/extinction dynamics. We argue that predictions of equilibrium species richness will be most accurate
when considering an eco-evolutionary island biogeography that integrates the traditionally considered effects of patch connectivity and area (Figure 4A) with their effects via (mal)adaptation (Figure 4B,C). In any given metacommunity, the traditional and (mal)adaptation-mediated effects of connectivity and area on species richness might operate alone or in concert. If both pathways are operating simultaneously, they very well might have effects on species richness of contrasting magnitude, and might even have effects in contrasting directions. For example, in any scenario where connectivity increases maladaptive gene flow, and maladaptation leads to reduced species richness, traditional and (mal)adaptation-mediated effects of connectivity on species richness should oppose one another. We therefore argue that it is valuable (and possibly essential) to evaluate the effects of both mechanisms because eco-evolutionary island biogeography can yield predictions of species richness that equal, exacerbate (Figure 4B), nullify, or even invert (Figure 4C) traditionally recognized effects of patch connectivity and area. Two empirical examples from recent eco-evolutionary research highlight these effects and support their general application. Timema cristinae stick insects Gene flow between T. cristinae populations locally adapted to different host-plant species causes populations to become poorly camouflaged (maladapted) on their resident host plants [31]. Thus, connectivity among populations on different host-plant species strongly influences the degree of camouflage maladaptation [51]. Recent experiments show that camouflage maladaptation has the ecological consequence of decreasing the species richness of cohabitating arthropods, such as caterpillars and beetles, because avian predators that eat stick insects are attracted to host-plant patches harboring maladapted Timema [9]. In this scenario, the (mal)adaptation-mediated effect of connectivity is to decrease equilibrium species richness, opposing the traditionally predicted effect of connectivity (Figure 4A). If the (mal)adaptation-mediated effect outweighs the traditional effect, equilibrium arthropod species richness could actually be lower on well-connected patches in nature (Figure 4C). Further landscape-scale experimentation and observation will be necessary to determine whether natural patterns reflect the findings of the experiments, and to compare the magnitude and direction of the multiple effects of connectivity. Threespine stickleback Gene flow between lake and stream populations of threespine stickleback (Gasterosteus aculeatus) causes maladaptation in foraging morphology [52], and some evidence suggests it might reduce stickleback population density [53]. In a stream habitat, where well-adapted stickleback feed on benthic macroinvertebates, maladaptive gene flow from lake populations should relax predation pressure by decreasing foraging efficiency on the benthos. This effect could increase successful macroinvertebrate colonization from nearby streams and reduce the likelihood of extinction, yielding higher equilibrium species richness for this guild. 157
Opinion
Trends in Ecology & Evolution March 2015, Vol. 30, No. 3
(A)
Rate of immigraon or exncon
Key:
Ex
nc
Isolated (I) Connected (C)
Im
m
on
igr
a
on
Species richness I
(B)
Isolated, classical Connected, classical Isolated, eco-evoluonary Connected, eco-evoluonary
Ex
on
n
a
c
igr
on
m
Im
Rate of immigraon or exncon
Key:
< C
Species richness Classical: Eco-evoluonary:
(C)
I I
C C
Isolated, classical Connected, classical Isolated, eco-evoluonary Connected, eco-evoluonary
Ex
on
n
c
a igr
on
m
Im
Rate of immigraon or exncon
Key:
<
